class: top, left, inverse, title-slide .title[ # Decomposing temporal changes of biodiversity to loss, survival, and recruitment of individuals: analysis of the North-American avifauna ] .author[ ### <br><br><br><br>François Leroy<br>Marta Jarzyna<br>Petr Keil ] .institute[ ### <br><br><br><br><br>Czech University of Life Sciences<br>Prague ] --- # Biodiversity * Multi-faceted (species richness, turnover, abundance) * Changes through time * Perturbed (mostly) by human activities -- <br> .center[ ## Abundance ] * the best metric to follow biodiversity state * overall decline of 69% (*acc.* Living Planet Index) .center[ `\(\Rightarrow\)` Marine Sharks and Rays: -71%<br><br> `\(\Rightarrow\)` Freshwater: -83%<br><br> `\(\Rightarrow\)` -3 billion North-American birds since 70's ] --- # Decomposing abundance change * Temporal change of abundance ( `\(\Delta N\)` ) results of `\(3\)` processes <br><br><br> .center[ <img src="data:image/png;base64,#images/deltaN_concept1.png" width="100%" /> ] <br> `$$\Rightarrow \Delta N = N_{t1} - N_{t0} = R - L$$` --- # Decomposing growth rate * Growth rate: per individual average change <br> .center[ <img src="data:image/png;base64,#images/deltaN_concept2.png" width="100%" /> ] .pull-left[ * Loss rate = individual probability of dying/emigrating * Loss rate `\(= \frac{L}{N_{t0}}\)` ] .pull-right[ * Recruitment rate = individual probability of producing an individual * Recruitment rate `\(= \frac{R}{N_{t0}}\)` ] -- .center[ `\(\Rightarrow\)` Growth rate = `\(\frac{\Delta N}{N_{t0}} = \frac{R}{N_{t0}} - \frac{L}{N_{t0}}\)` ] --- class: inverse, center, middle ## Problem Assessing the ongoing decrease of abundance does not tell us which process is driving the trend -- ### Increase in recruitment? Better reproductive success, better breeding habitat... -- #### Or ### Decrease of individual loss? Reduction of threats, better habitat quality... --- # Data - Breeding Bird Survey (BBS) .pull-left[ * Structured data on North American avifauna * From 1987 to 2021 ] .pull-right[ * More than 1000 routes * 564 species ] .center[ <img src="data:image/png;base64,#images/map_routes.jpg" width="70%" /> ] --- # Model - Dail & Madsen * Bayesian hierarchical model `$$N_{t+1} = Survival_{t+1} + Recruitment_{t+1}$$` .center[ #### with ] `$$Survival_{t+1} \sim Bin(N_t, \omega)$$` `$$Recruitment_{t+1} \sim Poisson(\gamma)$$` .footnote[ **Note:** `\(Loss_t = N_t - S_{t+1}\)` ] -- <br> * Accounting for detection probabilities, depending on meteorological covariates: <br> .center[ `\(\Rightarrow\)` **Sky condition, wind condition, temperature** & **time of the day**] --- class: inverse, center, middle ## We assessed ### 1) temporal change of abundance ( `\(\Delta N\)` ), recruitment ( `\(\Delta R\)` ) and loss ( `\(\Delta L\)` ) from 1987 and 2021 all across North America -- ### 2) temporal change of growth rate ( `\(\Delta N_r\)` ), recruitment rate ( `\(\Delta R_r\)` ) and loss rate ( `\(\Delta L_r\)` ) -- ### 3) which process, between recruitment (rate) and loss (rate), is driving the temporal change of abundance --- # Change in abundance, recruitment and loss -- .center[ <img src="data:image/png;base64,#images/map_ab.jpg" width="100%" /> ] --- # Change in abundance, recruitment and loss .center[ <img src="data:image/png;base64,#images/map_rec.jpg" width="100%" /> ] --- # Change in abundance, recruitment and loss .center[ <img src="data:image/png;base64,#images/map_loss.jpg" width="100%" /> ] --- # Change in abundance, recruitment and loss .center[ <img src="data:image/png;base64,#images/map_all_number.jpg" width="90%" /> ] --- # Recruitment vs. Loss .center[ <img src="data:image/png;base64,#images/recVsLoss1.png" width="75%" /> ] --- # Recruitment vs. Loss .center[ <img src="data:image/png;base64,#images/recVsLoss2.png" width="75%" /> ] --- # Recruitment vs. Loss .center[ <img src="data:image/png;base64,#images/recVsLoss3.png" width="75%" /> ] --- # Recruitment vs. Loss .center[ <img src="data:image/png;base64,#images/recVsLoss4.png" width="75%" /> ] --- # Recruitment vs. Loss .center[ <img src="data:image/png;base64,#images/recVsLoss5.png" width="75%" /> ] --- # Recruitment vs. Loss .center[ <img src="data:image/png;base64,#images/recVsLoss6.png" width="75%" /> ] --- # Recruitment vs. Loss .center[ <img src="data:image/png;base64,#images/recVsLoss7.png" width="75%" /> ] --- # Recruitment vs. Loss .center[ <img src="data:image/png;base64,#images/recVsLoss_ab.png" width="75%" /> ] --- # Recruitment vs. Loss .center[ <img src="data:image/png;base64,#images/recVsLoss9.png" width="75%" /> ] --- # Recruitment vs. Loss .center[ <img src="data:image/png;base64,#images/recVsLoss_colors.png" width="75%" /> ]